Sterile Neutrinos. Patrick Huber. Center for Neutrino Physics Virginia Tech

Transcription

1 Sterile Neutrinos Patrick Huber Center for Neutrino Physics Virginia Tech XIII Workshop on Electron-Nucleus Scattering June 23-27, 2014, Marciana Marina, Isola d Elba, Italy P. Huber p. 1

2 Outline The reactor anomaly Other evidence Future efforts P. Huber p. 2

3 Neutrinos from fission 235 U +n X 1 +X 2 +2n with average masses of X 1 of about A=94 andx 2 of about A=140. X 1 and X 2 have together 142 neutrons. The stable nuclei with A=94 and A=140 are 94 40Zr and Ce, which together have only 136 neutrons. Thus 6β-decays will occur, yielding 6 ν e. About 2 will be above inverse β-decay threshold. The problem is to determine how many exactly are above inverse β-decay threshold. P. Huber p. 3

4 Beta decay theory In Fermi theory, theβ-spectrum of electrons is given by Y(E e ) = k(e 0,Z) Ee 2 m 2 ee e (E 0 E e ) 2 F(E e,z) with the Coulomb correction for a point-like nucleus F(E e,z) = ψ(r = 0) 2 and k being a normalization factor P. Huber p. 4

5 Beta decay theory In Fermi theory, the spectrum of massless neutrinos is obtained from E ν = E 0 E e In reality there are many corrections: finite nuclear size, radiative corrections, screening effects, induced currents,... which in principle can be computed for allowed decays but not for forbidden ones. There is a sizable fraction of around 40% of all neutrinos coming from forbidden decays, essentially for reasons of combinatorics. P. Huber p. 5

6 Neutrinos from fission For a single branch energy conservation implies a one-to-one correspondence betweenβ and ν spectrum. However, here there are about 500 nuclei and individual β-branches involved; many are far away from stability. Directβ spectroscopy of single nuclei never will be complete, and even then one has to untangle the various branches γ spectroscopy yields energy levels and branching fractions, but with limitations, cf. pandemonium effect P. Huber p. 6

7 β-spectrum from fission 235 U foil inside the High Flux Reactor at ILL Electron spectroscopy with a magnetic spectrometer Same method used for 239 Pu and 241 Pu For 238 U reliance on the theory small contribution to overall neutrino spectrum Schreckenbach, et al P. Huber p. 7

8 Extraction of ν-spectrum The totalβ-spectrum is a sum of all decay branches N β (E e ) = de 0 N β (E e,e 0 ; Z)η(E 0 ). with Z effective nuclear charge andη(e 0 ), the underlying distribution of endpoints This is a so called Fredholm integral equation of the first kind mathematically ill-posed, i.e. solutions tend to oscillate, needs regulator. This approach is the basis for virtual branches Schreckenbach et al., 1982, 1985, 1989 and is used in the modern calculations as well Mueller et al. 2011, Huber 2011 P. Huber p. 8

9 Virtual branches 10 2 E MeV,Η E MeV,Η E MeV,Η counts per bin counts per bin counts per bin E e MeV E e MeV E e MeV 1 fit an allowedβ-spectrum with free normalizationη and endpoint energye 0 the lastsdata points 2 delete the lastsdata points 3 subtract the fitted spectrum from the data 4 goto 1 Invert each virtual branch using energy conservation into a neutrino spectrum and add them all. e.g. Vogel, 2007 P. Huber p. 9

10 Corrections toβ-shape There are numerous correction to the β-spectrum Size of correction WM weak magnetism G Ν QED radiative correction Cweak finite size L 0 QED finite size Sscreening by selectrons C L 0 WM G Ν S E Ν MeV Many of these correction depend on the nuclear chargez, butz is not determined by theβ-spectrum measurement nuclear databases. For forbidden decays many of these corrections are not known potentially large uncertainty. P. Huber p. 10

11 Reactor antineutrino fluxes 0.15 ΦΦ ILL Φ ILL our result ILL inversion simpleβshape E Ν MeV Shift with respect to ILL results, due to a) different effective nuclear charge distribution b) branch-by-branch application of shape corrections P. Huber p. 11

12 Improving a priori calculations ν e /MeV/fission ν e /MeV/fission ν e /MeV/fission Ratio Summation method / 1.2 P. Huber Energy (MeV) 239 Pu 238 U Ratio Summation method / 1.2 P. Huber Energy (MeV) 241 Pu 235 U Updated β-feeding functions from total absorption γ spectroscopy (safe from pandemonium) for the isotopes: 102,104,105,106,107 Tc, 105 Mo and 102 Nb Energy (MeV) Energy (MeV) Ratio Summation method / 1.2 P. Huber Energy (MeV) Energy (MeV) Still a 10-20% discrepancy with the measured total β- spectra. Fallot et al., 2012 P. Huber p. 12

13 Weak magnetism &β-spectra Nuclear structure effects can be summarized by the use of appropriate form factorsf N X. The weak magnetic nuclear,fm N form factor by virtue of CVC is given in terms of the analog EM form factor as FM(0) N = 2µ(0) The effect on the β-decay spectrum (in allowed decays) is given by 1+δ WM W M F N M (0) F N A (0) W P. Huber p. 13

14 What is the value of δ WM? Three ways to determineδ WM impulse approximation universal value 0.5%MeV 1 using CVC F M from analog M1γ-decay width, F A fromft value direct measurement in β-spectrum only very few, light nuclei have been studied. In those cases the CVC predictions are confirmed within (sizable) errors. Overall, good agreement with naive expectations except for handful of nuclei hint to non-trivial nuclear structure effects P. Huber p. 14

15 WM in forbidden decays k(e ν )/k(e ν ) original Treat all transitions as allowed GT Treat all non-unique forbidden transitions as [Σ,r] 0- Treat all non-unique forbidden transitions as [Σ,r] 1- Treat all non-unique forbidden transitions as [Σ,r] E ν (MeV) Approximate upper bound for the flux error due to forbidden decays. Hayes et. al, 2013 point out that in forbidden decays a mixture of different operators are involved, and that while for many of the individual operators the corrections can be computed, the relative contribution of each operator is generally unknown. My interpretation: it is the WM which is the leading cause for the large combined uncertainty they find. P. Huber p. 15

16 Recent data Measurement of 238 U beta spectrum Total flux from Daya Bay AD2 AD3 AD4 AD5 AD6 Model aryad1 U : 238 U : 239 Pu : 241 Pu : : : Y 0 (cm 2 GW 1 day 1 ) σ f (cm 2 fission 1 ) Data / Prediction (Huber+Mueller) ± Data / Prediction (ILL+Vogel) ± inary RENO bump Haag et al % agreement with a priori calculations, as expected. Data / Prediction Reference Model: Huber (3 isotopes) + Mueller ( 238 U) Previous average R = (exp.) Previous data Daya Bay World Average 1-σ Exp. Unc. 1-σ Flux Unc Distance (m) Daya Bay, Nu 2014 Daya Bay confirms the magnitude of the RAA, but does not change significance 19 RENO, Nu 2014 Double Chooz sees similar feature Origin? P. Huber p. 16

17 The reactor anomaly ROVNO88_3S 18.2 m ROVNO88_2S 25.2 m ROVNO88_1S 18.2 m ROVNO88_2I 18.0 m ROVNO88_1I 18.0 m SRP-II 23.8 m SRP-I 18.2 m Krasnoyarsk-III 57.3 m Krasnoyarsk-II 92.3 m Krasnoyarsk-I 33.0 m ILL 8.76 m Goesgen-III 65.0 m Goesgen-II 46.0 m Goesgen-I 38.0 m Bugey m Bugey m Bugey-3/ m ROVNO m 0.92X±0.01X± X±0.01X± X±0.01X± X±0.01X± X±0.01X± X±0.01X± X±0.01X± X±0.01X± X±0.18X± X±0.03X± X±0.06X± X±0.04X± X±0.02X± X±0.02X± X±0.11X± X±0.01X± X±0.00X± X±0.02X±0.03 Bugey-3/ m 0.93X±0.00X±0.03 τ =881.5s n PDG2010 Average 0.927X ± The increase in predicted neutrino fluxes, triggered a re-analysis of existing reactor data And this was found by Mueller et al., 2011, 2012 where are all the neutrinos gone? ν Measured / ν Expected, NEW P. Huber p. 17

18 Reactor anomaly Distance to Reactor (m) N OBS /(N EXP ) pred,new ROVNO88 3S ROVNO88 1S SRP I ILL Bugey 3 Bugey 4 ROVNO88 2I ROVNO91 ROVNO88 1I SRP II ROVNO88 2S Krasnoyarsk I Bugey 3 Goesgen I Goesgen II Krasnoyarsk III Goesgen III Krasnoyarsk II Bugey 3 PaloVerde CHOOZ DoubleCHOOZ Nucifer (2012) 6% deficit of νe from nuclear reactors at short distances 3% increase in reactor neutrino fluxes decrease in neutron lifetime inclusion of long-lived isotopes (non-equilibrium correction) P. Huber p. 18

19 Non-equilibrium corrections only 2 dozen isotopes with t 1/2 > 12 h above inverse β-decay threshold Mueller, et al., 2011 Extra shift due to long-lived isotopes a) small nuclear physics uncertainty inβ-decay b) depends on detailed fuel history P. Huber p. 19

20 Neutron lifetime 910 lifetime data from Wietfieldt & Greene, Rev. Mod. Phys neutron lifetimes range used in past reactor analyses PDG 2012 Mention et al IBD cross section change year P. Huber p. 20

21 Gallium anomaly 25% deficit of ν e from radioactive sources at short distances Effect depends on nuclear matrix elements This measurement was intended as a calibration is R a physics measurement or a calibration constant? P. Huber p. 21

22 Nuclear matrix elements P. Huber p. 22

23 Where are all the neutrinos? A simple explanation for all these results is given by neutrino oscillation P(ν e ν e ) = 1 sin 2 2θsin 2 m2 L 4E explains both the reactor and gallium results. L/E reactor = m/MeV L/E gallium 5m/MeV L/E = 5m/MeV corresponds to m eV 2 P. Huber p. 23

24 Why sterile? We need a m 2 of order ev 2, but we already have measured two other m 2 to be around ev 2 and ev 2. Given m ij = m 2 i m2 j, 3 different values of m2 require at least 4 different values for m we need 4 or more neutrinos The number of neutrinos coupling to thez is N ν = ± Any extra light neutrinos do not participate in weak interactions they have NO Standard Model interactions at all likely portal to hidden sector. P. Huber p. 24

25 Sterile oscillation In general, in a 3+N sterile neutrino oscillation model one finds that the energy averaged probabilities obey the following inequality P(ν µ ν e ) 4[1 P(ν e ν e )][1 P(ν µ ν µ )] independent of CP transformations. Therefore, a stringent test of the model is to measure P(ν µ ν e ) appearance P( ν µ ν e ) appearance P(ν µ ν µ ) or P( ν µ ν µ ) disappearance P(ν e ν e ) orp( ν e ν e ) disappearance P. Huber p. 25

26 LSND and MiniBooNE P( ν µ ν e ) The L/E values correspond to a m eV 2 P. Huber p. 26

27 Disappearance constraints m , 99, CL, 2 dof disappearance appearance m null results combined LSNDreactors Ga MB app null results disappearance No effects in - atmospheric - Bugey - CDHS - MINOS -... sin 2 2Θ Μe CL, 2 dof sin 2 2Θ Μe null results appearance Resolution will require new experiments, both for appearance and disappearance Figure from arxiv P. Huber p. 27

28 Astrophysics Effective neutrino species from Planck Abe, et al., 2013 N eff = 3.36±0.34 but, Planck also finds the Hubble parameter at H 0 = 67.4±1.4kms 1 Mpc 1 which is about 2.6σ below the value found by the Hubble Space Telescope. Combining HST and Planck yields N eff = 3.62±0.25 quoting from the Planck paper It is up to the reader to decide how to interpret such results, but it is simplistic to assume that all astrophysical data sets have accurately quantified estimates of systematic errors. We have there- P. Huber p. 28

29 Future reactor efforts There is number of planned reactor experiments Nucifer France research liquid (Gd) data taking DANSS Russia PWR plastic (Gd) under construction Stereo France research liquid (Gd) approved SoLid Belgium research plastic ( 6 Li) prototype Prospect USA research liquid R&D Poseidon Russia research liquid (Gd) R&D Neutrino 4 ussia researc liquid (Gd) under construction Hanaro Korea TBD liquid R&D There is a proposal to measure relevant β-spectra from a Cf spontaneous fission source. P. Huber p. 29

30 Future reactor sensitivity [ev 2 2 m ] RAA best fit RAA 99% CL RAA 95% CL SoLid 95% CL SoLid 99% CL SoLid 5σ 2 fit Rate + Shape Lhuillier, Nu sin 2 (2θ) Shown: Solid sensitivity Most experiments, on paper have similar sensitivity Challenge: backgrounds in close proximity to reactor P. Huber p. 30

31 Future source experiments 51 Cr EC 40 d 750 kev (90%) Th. n on 50 Cr 144 Ce β 411 d <2.997 MeV Fission product 8 Li β 868 ms <12.9 MeV ISODAR Challenges Intensity Shielding for Cr and Ce, detection threshold For Ce and Cr, liquid scintillator, Li may also work with Gd-doped water Cerenkov detectors P. Huber p. 31

32 Future source experiments 2 m 41 ev Liquid Xedepleted Liquid Xenatural SOX5 MCi sin 2 2Θ 14 Coloma, PH, Link, 2014 Combine a 5 MCi 51 Cr source with either a 100 t liquid scintillator detector (Borexino SOX) or a 6 t liquid Xe detector (LZ-like), which also has a very good sensitivity to neutrino magnetic moments Ce sensitivity for > 0.2eV 2 similar for 100 times less activity P. Huber p. 32

33 ISODAR Online production using spallation neutrons from a 60 MeV proton beam ν e ν x? Observed/predicted ν e ν e 5 yrs 5 KamLAND L/E (m/mev ) 820,000 IBD events in 5 years at KamLAND (16 m baseline to center of detector) Spitz, Nu 2014 Spitz, Nu 2014 P. Huber p. 33

34 OSC-SNS Redo LSND at the SNS, most direct test of LSND possible P. Huber p. 34

35 nustorm Neutrinos from stored muons no cooling, no acceleration FIG. 1. A schematic of the storage ring configuration. Pions Sensitivity is unprecedented and also allows high precision disappearance tests m 2 [ev 2 ] σ, 1% Sys 99% C.L., 1% Sys 10σ, 5% Sys 99% C.L., 5% Sys 99% C.L. Fit to Evid. 99% C.L. Fit to App. 99% C.L. Icarus sin 2 2θ eµ P. Huber p. 35 D. A

36 Summary Reactors are complex sources the reactor anomaly may be a true deficit in neutrinos or just a result of the complexity of the source Other hints for sterile neutrinos, but tension with disappearance signals Many experiments possible on a 5 year time scale Discovery of sterile neutrinos would we paradigm shifting P. Huber p. 36